[tvkalvas]

Hi!

I have been following the headphone diy community for a while and I can't find out why so many headphone amplifiers are using a virtual ground instead of real dual regulated power supply with a real ground? For PPA for example a large fraction of costs could be cut by having only two channels and a real ground power supply.

Just trying to understand

 

[NelsonVandal]

Quote:

Originally Posted by tvkalvas /img/forum/go_quote.gif Hi! I have been following the headphone diy community for a while and I can't find out why so many headphone amplifiers are using a virtual ground instead of real dual regulated power supply with a real ground? For PPA for example a large fraction of costs could be cut by having only two channels and a real ground power supply. Just trying to understand

Do you mean "real ground" as in ground to negative rail?

Or do you mean "real ground" as in a common ground for both the headphones return and small signal return, not using an active ground channel for the headphones return currents?

 

[cobaltmute]

I think you mean ground channel versus virtual ground.

Virtual grounds, to my pea-brain, create a zero reference for the signal when dealing with a single supply such as a 9v battery.

The ground channel is the third channel coming from source to output. You have left/right/ground. To my understanding, by treating the ground channel with as much care as the left and right, you end with a more "black" signal overall. Black meaning clean in that there is no hiss, no crosstalk etc.

 

[tvkalvas]

I mean that in classical power and pre amplifier designs there is only one ground. This ground is halfway between positive and negative rails (+15V and -15V in typical preamps) and both input and output negative terminals/return currents are connected to this ground. If a headphone amplifier was built like this, the input ground and headphones return current would come to this same place.

 

[nikongod]

The more expensive stuff does use +/- dual regulated power supplies. I think a LARGE part just comes down to cost to build. With a PPA / M^3 / Pimeta ETC you only NEED 1 floating power supply & a rail-splitter which is much less expensive to build. You could (theoretically) build them with a dual-regulated PSU if you wanted. Even if you redesigned something like the steps to be a dual PS with one transformer secondary/rail to save on the cost of the most expensive part, you would still have to duplicate EVERYTHING after that.

As far as active ground channels: they can reduce distortions.

 

[tvkalvas]

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif Virtual grounds, to my pea-brain, create a zero reference for the signal when dealing with a single supply such as a 9v battery.

Sure... this I understand. But if you are working with a line powered device. Then it doesn't make sense to me.

Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif The ground channel is the third channel coming from source to output. You have left/right/ground. To my understanding, by treating the ground channel with as much care as the left and right, you end with a more "black" signal overall. Black meaning clean in that there is no hiss, no crosstalk etc.

Thinking about the technical side of the design / circuit theory, I cannot see how using an active ground channel would make a better amplifier. Star-shaped (one point) grounding should keep the influence of ground wire currents to minimum. Maybe the DIY gurus have made tests comparing traditional designs and active ground designs?

 

[Ferrari]

Quote:

Originally Posted by tvkalvas /img/forum/go_quote.gif ... Thinking about the technical side of the design / circuit theory, I cannot see how using an active ground channel would make a better amplifier. Star-shaped (one point) grounding should keep the influence of ground wire currents to minimum. Maybe the DIY gurus have made tests comparing traditional designs and active ground designs?

Conventional 2-channel "passive ground" amplifier differs from 3-channel "active ground" amplifier in the way how the return current from the drivers (from the speakers in the headphone) is handled. A conventional 2-channel "passive ground" amplifier just dumped the return current from the drivers into the "passive ground" => ground polutions.

As opposed to 2-channel "passive ground" amplifiers, in a 3-channel "active ground" amplifier, the "ground" wire of the headphone is driven by a third channel of the same topology as the left and right channels (thus 3 amplifiers of the same design). The "active ground" (ground channel amplifier) sinks or sources the return current from the drivers. Ground polutions (common in 2-channel "passive ground") can be avoided while improving the linearity and stereo crosstalk.

 

[Beefy]

Quote:

Originally Posted by tvkalvas /img/forum/go_quote.gif Sure... this I understand. But if you are working with a line powered device. Then it doesn't make sense to me.

I think your cost/benefit is a bit off. To get a top-shelf dual rail power supply costs a lot more than a single rail with a virtual ground.

Especially considering it is only the voltage gain OPAMPs using the +/GND/-, and this doesn't need huge amounts of current. The high current output buffers run quite happily over just the single +/GND rail.

Quote:

Thinking about the technical side of the design / circuit theory, I cannot see how using an active ground channel would make a better amplifier. Star-shaped (one point) grounding should keep the influence of ground wire currents to minimum. Maybe the DIY gurus have made tests comparing traditional designs and active ground designs?

Because it stops the headphone ground from directly seeing the power supply ground. It puts a large buffer in between that keeps the ground steady and clean.

 

[NelsonVandal]

The grounding affects the sound just as much as the left/right channel, so care has to be taken. When using "normal" ground, the sound will depend on the quality of rails-to-ground capacitors and/or a regulated power supply with low output impedance (if there is any). With an active ground you'll get a low output impedance for the whole frequency spectrum, a better channel separation and in my oppinion a better/truer stereo image.

I'm not sure if there's any improvement from separating the small signal ground from headphones return. I can't really say I can hear any difference.

 

[majkel]

Agreed with NelsonVandal. Many dual supply rail amps have problems with the soundstage coherence, which does not happen with tube or virtual ground SS amps. Tube amps don't use dual supplies either.

 

[tvkalvas]

Ok guys... you got me convinced. I will go for the real PPA v2 then, not a two channel modification of it. Also it is easier for me because I can use the pcb from Tangent's shop and I don't have to mess with acids.

Then I have to attack the next problem... sourcing all the parts. But that will go in another thread/search.

Thank you all for your feedback.

 

[NelsonVandal]

One good thing with virtual ground, active ground channel and classA is you don't have to buy large and expensive capacitors, since the amp can be DC-coupled and the power caps play a minor role. No more grainy-sounding-capacitor-neurosis. There's enough trouble with sore-ear-opamp-depression and trying-to-find-exotic-transistors blues.

Another good thing with the active ground channel topology is that you can blend opamps to get the sound you want, a sound that suit your ears and phones.

 

[nikongod]

you are seriously confusing active grounds, rail splitters, and virtual grounds.
Quote:

Originally Posted by majkel /img/forum/go_quote.gif Agreed with NelsonVandal. Many dual supply rail amps have problems with the soundstage coherence, which does not happen with tube or virtual ground SS amps. Tube amps don't use dual supplies either.

lets make a short list of dual-supply-rail amps.
the beta 22, the vast majority of the gilmore designs (most of his tube stuff and all of his solid state) and as the OP mentioned the vast majority of power amps in production.
Quote:

Originally Posted by NelsonVandal /img/forum/go_quote.gif One good thing with virtual ground, ... since the amp can be DC-coupled

you can DC couple designs with "real grounds" and proper +/- supplies too. There is nothing prohibiting you from building an amp with a +/- supply centered on the ground pin at the wall, and running this as ground return OR an active ground "output" channel. The B22 with 3 channels is typically constructed this way, but a few people have built 2-channel b-22. there are VERY few 3-channel dyna-whatevers, and they still sound good.

 

[Pars]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif <snip> The B22 with 3 channels is typically constructed this way, but a few people have built 2-channel b-22. there are VERY few 3-channel dyna-whatevers, and they still sound good.

I don't know of any 3-channel Dyna-whatevers; they're either 2 channel or balanced (4-channel). Kevin Gilmore deplores rail splitters and active grounds. If you do some searches, you should be able to find some posts if interested.

 

[00940]

I once discussed the issue of an active ground on diyaudio... the answer was "why would you want to do that for an AC powered amplifier ?"

Arguments given against the active ground were that it would have twice the distortion and twice the output impedance of an amplifier with dual rails.

And about the arguments given in favor of the active ground, here were the answers:

- Removing the caps out of the equation: just use good regulators and almost no capacitance.
- Preventing contamination of signal ground: just use a decent grounding wiring scheme.
- Getting a tracking PS for free: not very useful with most amps.